Polyherbal phytosomal gel for enhanced topical delivery: design, optimization by central composite design, in vitro and ex-vivo evaluation

Abstract Annona squamosa (AS) and Cinnamomum tamala (CT) leaves have been used traditionally and have well reported various pharmacological activities which justify their topical use. But the majority of compounds present are phenolic compounds which due to their polar nature are unable to permeate through the skin. Therefore, polyherbal phytosomes of ethanolic extracts of AS and CT were combined in a ratio of 1:1 to get polyherbal Ethanolic extract (PHEE) and polyherbal phytosomes (PHP) were prepared and optimized using Response Surface methodology by applying Central Composite design (CCD) for studying the effect of drug soya lecithin ratio and speed of rotation on responses like vesicle size and entrapment efficiency. Followed by numerical optimization, PHP with vesicle size of 215.5 ± 0.45 nm and Entrapment efficiency of 85.24 ± 0.05% were obtained. These PHP were characterized with FTIR, DSC, SEM and XRD. The PHP showed % cumulative drug release of 84.22 ± 5.62%. For easy application on the skin these phytosomes were incorporated in Carbopol 934 gel matrix and evaluated for physical appearance, pH, viscosity and drug content. The ex vivo permeation of the phytosomal gel (PG) was compared with a conventional gel containing PHEE. A significant improvement in permeation was observed in case of PG with 36.62 ± 3.33 µg/cm2 of drug permeated in 24 hours whereas 17.06 ± 1.54 µg/cm2 through the conventional gel. The PHP and PG were found stable during stability studies for six months. Polyherbal phytosomal gel can be a good carrier for herbal extracts for topical delivery. GRAPHICAL ABSTRACT


Introduction
The traditional uses of Annona squamosa (AS) commonly known as Custard Apple as a medication have been documented in several different ways.Vermicidal and insecticidal components are found in leaves, seeds, and unripe fruits.Leaves are also utilized as an anthelmintic.For unhealthy ulcers, leaves are ground into a paste without the addition of water. [1]In various parts of India leaves of AS have been used as paste for curing wounds of cattle and also consumed with pepper for treating diabetes. [2]Various flavonoids have been separated from the leaves and evaluated for their antioxidant and hypoglycemic properties.The main flavonoids which were separated are quercetin-3-O-robinobioside, rutin, quercetin-3-O-b-D-glucoside, kaempferol-3-O-robinobioside, and kaempferol-3-O-rutinoside. Following the scavenging of DPPH free radicals, the isolated compounds demonstrated considerable antioxidant activity. [3]he essential oils of leaves were also investigated using the gas chromatography-mass spectrometry (GC-MS).
As it can now be said that both plants possess properties that support their use as topical dosage form for their antiinflammatory, antimicrobial and wound healing activity.According to several studies, combining herbs with different pharmacological properties and potencies in an ideal ratio might theoretically result in more positive outcomes than using the herbs separately.The term "polyherbalism" has historically been featured in the Ayurvedic literature "Sarangdhar Samhita," which dates back to 1300 A.D. [15] Plant combinations and mixed extracts are preferred over single ones in the ancient Indian medical system.Even while the active phytochemical components of specific plants have a well-established history, they are typically only found in trace amounts and are seldom enough to provide the desired therapeutic effects.Because of this, research has shown that combining these different plants with differing potencies might conceivably result in a better outcome than using them separately thus adding up their unique effects. [16]ost of the components present in the leaves of AS and CT are hydrophilic in nature and have high molecular weight which are too large to be absorbed by simple diffusion and have low permeability across cell membrane.Also, the essential oils present are sensitive to environmental conditions, leads to degradation and volatilization. [17]Therefore, making a complex between these extracts and phospholipid may overcome the problems of low bioavailability, permeability and stability.
PhytosomesV R , a patented technique created and marketed by Indena may create lipid-compatible molecular complexes by mixing standardized herbal extracts or polar components with phospholipids e.g., phosphatidylcholine (PC), which is obtained from soy beans.Complexes of phytoconstituents with phospholipids, or phytophospholipids, are helpful for enhancing the skin's absorption and bioavailability of active substances. [18]Phyto-phospholipid complexes are formed by the interactions between active components and the polar head of phospholipids.Phospholipid complexes, in which the phospholipids head group is anchored, can be developed by interactions between active ingredients and phospholipids, but the two long fatty acid chains are not involved in the formation of the complex.In order to create a lipophilic surface, the two long fatty acid chains can migrate and cause the encapsulation of the polar portion of complexes. [19]ccording to several research, phytophospholipid complexes can improve the absorption of herbal components through the topical route, thus increasing the bioavailability & lowering the required dosage.Consequently, it can greatly enhance therapeutic benefits.The oil-water partition coefficients and membrane permeability of active ingredients both considerably increased after the formation of phytophospholipid complexes.In comparison to free active constituents, phytophospholipids complexes are thus more easily absorbed and produce improved bioavailability.When applied topically, phytophospholipid complexes outperform an equivalent amount of the active component or extract in terms of biological activity.As a result, recent years have seen an upsurge in interest in the formulation of phytophospholipid complexes [20] Due to its lipophilicity, the stratum corneum, the outermost layer, makes it difficult to distribute this kind of medication.The phytosomal form of phytoconstituents helps to transfer the phytoconstituents into systemic circulation since it easily crosses this barrier.According to research investigations conducted over a 20-year period, phytosome technology is effective enough to serve as a new form of phytoconstituent carrier.The potency of phytosomal carriers allows them to be used topically to increase the bioavailability of phytoconstituents. [21]owadays QbD (quality by design) approach has been widely used to study the effect of various independent variables on the dependent variables.These variables may vary on the method employed for the preparation of phytosomes.In the light of above background, the research was associated with the design of polyherbal phytosomes (PHP) containing the extracts of leaves of AS and CT for topical delivery.Phytosomes were prepared by thin layer hydration method and optimized using the central composite design (CCD).The developed PHP were evaluated for vesicle size, zeta potential, entrapment efficiency, FTIR, DSC, and XRD study, in-vitro release study, etc.In order to make the application of PHP more convenient on the skin the optimized PHP were incorporated into a gel (Carbopol 934) and evaluated for certain characteristics parameters like physicochemical properties of gels, ex-vivo permeability, rheological behavior, texture analysis, and stability study.The findings explored that the incorporation of PHP into gel improved the penetration of herbal extracts through the skin and also results in ease of application. [22]

Materials
The dried leaves of AS were collected from local area in Bareilly Uttar Pradesh, India and CT were collected from local market at Gautam Budh Nagar Uttar Pradesh, India in February 2021.The leaves were authenticated at NISCAIR, New Delhi, India and a specimen was submitted (vouchers no: NISCAIR/RHMD/Consult/2020/3777-78-1 and 2).Soya Lecithin, Cholesterol, Methanol, Carbopol 934 and Standard Quercetin with purity 98%v/v were procured from Central Drug House New Delhi, India.All other chemicals were of analytical grade.

Preparation and phytochemical screening of plant extracts
The powdered leaves of AS and CT were shade dried and a coarse powder was prepared.This coarse powder was then used for Soxhlet Extraction using four different solvents in the ascending order of their polarity at their respective boiling points i.e., petroleum ether (60-80 ± 2 � C), chloroform (61.2 ± 2 � C), and ethanol (78.3 ± 2 � C) under continuous heat for about 72 hours or until becomes colorless and cold maceration was used for extraction with water.These extracts were then filtered using Whatman filter paper no. 1 and concentrated under reduced pressure in a rotary evaporator and all extracts were dried in vacuum and kept in a desiccator for further use.The phytochemical screening of the extracts was done for determination of flavonoids and phenolic compounds.

Estimation of quercetin
The bioactive marker quercetin was used for various quantitative purposes as already used in various earlier studies. [23]he estimation of quercetin using UV Visible spectroscopy was done by a method previously developed by Patil et al. [24] Standard curve of quercetin was prepared at a concentration of 2, 4, 6, 8, and 10 mg/mL in n-butanol, water and acetic acid in a ratio of 7:1:1.10 mg of ethanolic extracts of AS (ASEE) and CT (CTEE) each were dissolved in ethanol and diluted with n-butanol, water and acetic acid in a ratio of 7:1:1.The absorbance was determined at 256 nm using UV-visible Spectrophotometer (Systronics Double beam, Model 2202).

Formulation and optimization of polyherbal phytosomes
Phytosomes were prepared by the method of thin layer hydration as seen from the preliminary studies that this method yielded phytosomes with least particle size and highest entrapment efficiency. [24]ASEE and CTEE were taken in 1:1 ratio (100 mg each) to get a mixture of polyherbal ethanolic extracts (PHEE).Soya Lecithin and cholesterol were dissolved in Chloroform and methanol mixture in the ratio of 2:1.This solution was transferred to round bottom flask and evaporated in a rotary vacuum evaporator unit to get a thin film at 40 � C and 120 rpm.The film formed was kept in vacuum oven overnight for complete removal of the solvent.The dried film formed was then rehydrated with the help of phosphate buffer saline (PBS) pH 7.4 containing PHEE for 1 hour rotating at 80-160 rpm.The prepared dispersion of PHP was subjected to ultrasonication for 5 minutes to obtain phytosomes of desired size.The PHP dispersion was centrifuged at 16,000 rpm for 30 minutes using a cooling centrifuge and lyophilized.

Central composite design (CCD)
A CCD (Design-ExpertV R software, version 8.0.7.1, Stat-Ease, USA) was implemented to study the individual and combined effect of formulation variables on CQA (Critical Quality attributes).Initially, preliminary studies were conducted to decide the independent factors involved in phytosomal preparation using a thin-layer hydration method along the details available in previous literature. [25]Based on the literature and preliminary findings, Drug: Soya lecithin ratio (X1) and speed of rotation (X2) were chosen as independent parameters.The effect of decided independent variables (Drug: Soya lecithin ratio and speed of rotation) were studied on dependent variables such as vesicle size and entrapment efficiency.
During the optimization trial, these values for independent variables were varied between the extreme levels as 0, ± 1 and 1 ± a. (S1).The coded and real values of all 12 batches are depicted in S2.All 12 runs were prepared and analyzed for vesicle size and entrapment efficiency.The values of responses were again placed in the CCD of the design expert to find the projected values, experimental models, polynomial equations, and 3D surface plots/contour plots.The experimental models, viz.linear, 2nd order (2F1), and quadratic, were investigated to select the best-fitting model.

Validation of optimized model
Preparing an extra experimental formulation of phytosomes using the ideal values of X1 and X2 served to validate the CCD-generated model.The predicted values and the observed values were compared.In a sense, the experimental validation supported the design-generated quadratic model's robustness. [26]Using the equation shown below, the anticipated error or bias (%) between these two yields was determined.

Vesicular size and zeta potential.
The vesicle size and zeta potential were analyzed by using Zetasizer-1000HS (Malvern Instruments, UK) which was based on the theory of dynamic light scattering.Samples were diluted using distilled water.The samples were analyzed in a quartz cuvette at a scattering angle of 90 � . [27]Every batch was examined.
The mean and SD were then computed in triplicate.The impact of formulation factors on vesicular size was investigated.Zeta potential is considered as a significant parameter to forecast the stability of the particulate formulation.

Fourier transform infrared Spectrophotometry (FTIR).
In order to investigate the interaction between ASEE, CTEE, Soya Lecithin and Cholesterol and to determine the chemical stability and structure of the PHP complex, FTIR spectra of Soya Lecithin, Cholesterol, PHP complex and extracts were examined using Fourier transform infrared spectrophotometry (FTIR Spectrometer, BRUKER IFS-55, Switzerland).Using the potassium bromide (KBr) method, the IR spectra of ethanolic extracts, Soya Lecithin, and cholesterol, their complex and physical mixture were obtained.In order to make KBr pellets, a 1 mg sample was carefully mixed with 100 mg of KBr.Spectral scanning was carried out in the 400-4000 cm À 1 regions.

Differential scanning calorimetry (DSC).
The aluminum crimp cell was filled with ASEE, CTEE, Soya Lecithin, Cholesterol, physical mixture and PHP complex.The cell was heated from 0 to 350 � C in a nitrogen environment at a speed of 10 � C min À 1 .The peak transition onset temperatures were recorded by the help of an analyzer (Perkin Elmer Pyris-6 Version 4.0)

X-ray diffraction analysis (XRD)
. By using a PAN analytical X'PERT-PRO Powder X-ray diffractometer (Model: D8 ADVANCE, Bruker AXS, Inc., Madison, WI, USA), the XRD analysis was recorded.A 2� range in the spectrum's 10 � -80 � region was obtained when the sample was scanned at a rate of 0.3 seconds per degree at the CuKa radiation from the anode materials (1.54060).The X-ray generator was permitted to run at a tube current of 35 mA and a tube voltage of 40 kV.During the step scan mode, using a step time of 32.8 seconds, the scanning angle was varied between 3 � and 60 � .For ASEE, CTEE, Soya lecithin, cholesterol and optimized PHP formulation, separate spectra were run.

Scanning electron microscopy (SEM).
SEM was used to visualize PHP.SEM has been used to determine the complex's surface shape, appropriation estimation, and particle size.For an auxiliary electron emissive SEM (Nova Nano SEM 450), gold/palladium was used for coating of the samples for 120 seconds at 14 milliamperes under argon before being examined for the morphology at 15.0 kV.

In vitro diffusion study.
The Franz diffusion cell (Standard 15 mm) using dialysis membrane with a molecular weight cut off 10000 Da was used for study.The membrane was soaked in hot phosphate buffer pH 7.4 for 12 hours before conducting the diffusion study.The study was done to compare the release profile of conventional formulation (solution of 1:1 ratio of ASEE and CTEE) and prepared PHP dispersed in water (The quantity of extracts was fixed at 10 mg). 10 mL of both formulations was used for the studies.Phosphate buffer pH 7.4 (10 mL) was used as the release medium, and the investigation was conducted at a constant temperature of 37 � C. At predetermined intervals, such as 0, 0.25, 0.5, 1, 2, 4, 6, 12, and 24 hours samples (1 mL) were taken and refilled with the same volume of new media. [28]The obtained sample was examined for quercetin (bioactive marker) using the UV spectroscopy technique at 256 nm at each time point.

Drug release kinetics.
The drug release kinetics of the PHP formulation was studied by applying different kinetic models such as zero order, first order, Korsmeyer-Peppas and Higuchi model, models and then regression coefficients were determined.The different equations for various kinetic models are as follows.
Zero order : First order : ln where Q t is the percentage (%) drug released at time t, Q 0 is the initial value of Q t , t is the time, n is the diffusion release exponent, K 0 , K 1 , K t , and K H are the release coefficients corresponding to various kinetic models. [29]

Development of phytosomal gel (PG)
The optimized PHP formulation was incorporated into a gel.The gel was formulated using Carbopol 934 as a gelling agent (0.5, 1 and 1.5%w/w).Carbopol 934 in the required amount was dispersed in water and allowed to hydrate for 4-5 hours.Then Propylene glycol (5%w/w) was added to the Carbopol 934 dispersion.PHP with equivalents amount of 2% PHEE were first dispersed in water and then added to the hydrated gel.Triethanolamine was added to adjust the pH until a homogenous gel is obtained while stirring on a magnetic stirrer at 200 rpm.The prepared gel was kept for 6-8 hours to remove any entrapped air.Another gel containing only PHEE was also prepared by the same procedure.

Optimization of the phytosomal gel (PG).
The prepared gels were evaluated for their physical examination like pH, color, appearance, consistency and viscosity.The drug content of the gels was also determined.For this purpose, 1 gm gel was dissolved properly in ethanol.The resultant solution was used for determination of absorbance using UV-visible spectroscopy for estimation of drug content in terms of quercetin at 256 nm.All the readings were recorded in triplicate (mean ± S.D.)

Ex vivo permeation study.
The ex vivo permeation study was performed using Franz diffusion cell with a diffusion area of 1.77 cm 2 .Goat skin was obtained from a slaughter house.The skin was washed properly and hair were removed.It was then placed between two compartments, that is, donor and receptor with dermal layer toward the downward direction.Phosphate buffer 7.4 was used as the permeation medium.The donor side was filled with 1 gm of the Investigational gel (PHEE in gel and PHP in gel separately).The Franz diffusion cells was maintained at 200 ± 5 rpm at 37 � C. At predetermined intervals, such as 0, 0.25, 0.5, 1, 2, 4, 6, 12, and 24 hours samples (1 mL) were taken and refilled with the same volume of new media.The obtained sample was examined for quercetin (bioactive marker) using the UV spectroscopy technique at each time point.
A graph with cumulative amount of drug permeated per unit area (mg/cm 2 ) at the y-axis and time at the x-axis was plotted.With the help of this graph the permeation profiles were determined.The steady state flux (J ss ) of drug was calculated by applying the Linear regression analysis.The permeability co-efficient (K p ) of drug, across the stratum corneum, was calculated using the equation below. [30] where, C is the initial concentration of the drug in the donor compartment.The Enhancement ratio (ER) which describes the enhancement in penetration of drug through the skin was calculated by using the equation: The spread test fixture's cone was filled with the prepared gel, which was then given 15 minutes to set at room temperature.For analysis a smooth and flat surface is required.This smooth surface also prevents early test triggers.Therefore, extra formulation was removed by scrapping from the cone holder.The probe (TA10) was set to move into the gel at a pretest speed, a test speed and a return speed of 1, 0.5, and 0.5 mm/s respectively in texture profile analysis mode.Texture Pro CTv1.2 (Middleboro, MA, USA) software was used to gather the data and analyze it in order to identify several mechanical performance parameters such as hardness and adhesiveness.

Stability studies of PHP and PG
Stability studies were conducted as per ICH guidelines ICH Guidelines, Q1A (R2) 2003. [31]The optimized PHP formulation was transferred in to a clear glass vial and sealed.Then it was stored at three different conditions at low temperature (5 ± 3 � C), at 25 ± 2 � C/60 ± 5% relative humidity (RH) and at accelerated conditions (40 ± 2 � C/75 ± 5% RH) for 6 months.The samples were withdrawn at specified intervals of time, that is, 0, 1, 3, and 6 months and were evaluated for vesicle size and entrapment efficiency.The PG was kept in wide mouth glass bottles under the same above-mentioned conditions and evaluated for physical appearance and drug content.

Data analysis
The results were expressed as mean ± S.D. The statistically significant differences were detected using two-tailed and unpaired student's t-test.The level of significance was kept as P < 0.05.

Phytochemical Screening
Based on the phytochemical screening of different extracts of AS and CT leaves.It was found that the ethanolic extracts of both were found to be rich in various phytoconstituents so the ethanolic extracts of AS and CT leaves were taken in 1:1 ratio for preparation of phytosomes.

Quercetin content in ASEE and CTEE
The Calibration curve of quercetin in n-butanol, water and acetic acid in a ratio of 7:1:1 was found to be linear at a concentration range of 2-10 mg/mL (n ¼ 3) with the value of correlation coefficient as 0.998.The content as determined by UV Spectrophotometry was found to be 30.27± 0.155 mg/10 mg and 59.73 ± 0.155 mg/10 mg of the ASEE and CTEE respectively using standard curve of quercetin.Therefore 20 mg of PHEE contained 90 mg of quercetin.

Formulation and optimization of polyherbal phytosomes
The thin layer hydration method was used for formulating PHP of ASEE and CTEE (1:1) using Soya Lecithin and Cholesterol in organic solvents chloroform and methanol mixture in the ratio of 2:1.On the basis of preliminary studies it was found that speed of rotation and drug soya lecithin ratio had a significant influence on entrapment efficiency and vesicle size so, they were selected as independent variables.These variables were optimized through Design-ExpertV R software.

Optimization by CCD
As per CCD, 12 formulation batches were prepared as shown in Table 1.The responses studied were vesicle size (Y1) and entrapment efficiency (Y2).All the experiments were performed in triplicate.The data when filled in the design expert software provided polynomial equations.The data was optimized using ANOVA in the software.The software suggested various models which were evaluated in terms of statistically significant coefficients and R 2 values.The relationship between the independent variables and responses was assessed by using 3-D surface and contour plots.Highest entrapment efficiency (Y2), and smallest value of vesicle size (Y1) were kept as the criterion for selection of optimum formulations.A comparison between the resulting observed responses and the predicted responses was done.
The percent error was calculated and linear regression plots among actual and predicted responses were plotted.After the optimization the characterization of PHP was done.
As can be seen from the response surface parameters of PHP formulation the vesicle size was found in the range from 198.2 to 241.3 nm and the entrapment efficiency between 71.85% and 87.17%.By the end of different formulation trials, the following equations for vesicle size and entrapment efficiency were obtained.
The second-order polynomial equation providing the relationship of the response of vesicle size (Y1) is given below: Y1 ¼ þ222:78 þ 6:75 X1 À 11:79 X2À 3:95 X1X2 À 2:57 X1 2 À 5:22 X2 2 The F-value provided by the model was found to be 521.73 which suggested that that the model was significant (p < 0.0001).The Lack of fit value was observed as 2.79 suggested that Lack of fit was not significant (p ¼ 0.2508).In this case X1, X2, X1X2, and X1 2 and X2 2 are significant model terms.Positive coefficients of X1 in the equation recommended the synergistic effect of formulation variables on vesicle size while negative coefficients of X2, X1X2 and X1 2 and X2 2 recommended the antagonistic effect on the vesicle size.As seen from Table 1 it is found as the ratio drug: soya lecthin was increased form 1:1 to 1:3 the vesicle size was gradullay increased. [32]The reason behind this may be the excessive availaibilty of soya lecithin as compared to phytoconstituent molecules which come in contact during the formulation of phytosomes.There might be more collisions which may have lead to an increase in chances of agglomeration and the size of vesicles.
At a contant drug soya lecithin ration as seen in the formulations F1, F3 and F7, F8 increase in speed of roration caused decrease in the particle size.If we look at formulation F7 and F9 with drug: soya lecthin of 1:2 and speed of rotation of 63.43 and 120 rpm respectivley there was not a significant decrease in particle size but when the speed of rotation was increased to 176.57rpm as in case of F8 there was a significant decrease in partcle size. [33]These all variations in vesicle size due to the independent variables are well exhibited by the 3D response surface and contour plots in Figure 1a, b.The entrapment efficiency of the prepared polyherbal phytosomes were determined in triplicate and the results can be observed in Table 1.The second-order polynomial equation providing the relationship of the response of particle size (Y2) is given below: The F value provided by the model was found to be 269.83which suggested that that the model was significant (p < 0.0001).The Lack of fit value was observed as 1.08 suggested that Lack of fit was not significant (p ¼ 0.2508).In this case X1, X2, X1X2 and X1 2 and X2 2 are significant model terms.Positive coefficients of X1 and X1X2 in the equation recommended the synergistic effect of formulation variables on particle size while negative coefficients of X2, X1 2 and X2 2 recommended the antagonistic effect on the entrapment efficiency.As seen from the table it is found as the ratio drug: soya lecthin was increased form 1:1 to 1:3 the entrapement efficiency was gradullay increased (batch F1-F2, F3-F4) keeping the speed of rotation constant.This may be due the presence of more amount of Soya lecithin for encapsulation. [34]On further increasing the drug to soya lecithin ratio the entrapment efficiency did not increase significantly.This may have occurred due to less amount of drug available to form complex with more amount of soya lecithin.
The inference from the response surface plot suggest that on increasing the speed of rotation the entrapement efficiency increases up to a certain point and on further enhancement in the speed it decreased (batch F7-F9) .The higher speed may have cause rupturing of vesicles and thus loss of entrapped drug. [35]These all variations in entrapement efficiency due to the independent variables are well exhibited by the 3D response surface and contour plots in Figure 2a, b.The Summary of Results of Regression Analysis for Responses Y1 and Y2 are given in S3.The comparative graphs between the actual and predicted responses can be seen in S4.
The optimization of the experimental trial was done by applying numerical optimization technique.As per this technique, constraints were applied to both independent and dependent variables to get the optimized and most desired formulation.Various constraints were applied with lower limit of 1:1 and higher limit of 1:3 in case of drug soya lecithin ratio and lower limit of 80 rpm and higher limit of 160 rpm in case of speed of rotation.The goal was  minimum vesicle size and maximum entrapment efficiency.A new formulation with the desired results was developed utilizing a numerical optimization technique and the desirability approach.On the basis of achieving the highest percent entrapment efficiency and the smallest vesicle size, the best formulation was chosen.Following the study, one potential answer for the experimental trial was presented by the design, which demonstrated suggested values for independent variables and a desirability value near to 1 (S5).The desirability plot can be seen in S6.

Validation of model optimization
The validation of the model generated by CCD was done by preparing an extra formulation of phytosomes using the optimum values of X1 and X2.A comparison between the observed and predicted values was done.It was found that the experimental validation supported the design-generated quadratic model's robustness.By using the equation, the anticipated error or bias (%) between these two responses was determined.It was found to be less than 3%.Therefore, it can be concluded that the CCD successfully aided in optimizing the PHP formulation.The details of the validation can be seen in S7.

Entrapment efficiency
The Entrapment efficiency of the optimized batch followed by numerical optimization was found to be 85.24% ± 0.25 which was very close to the predicted value of 83.49% with a percent bias of 2.12 which is less than 3%.

Vesicular size and zeta potential
DLS method was used to determine the vesicle size and its distribution.The size and distribution of the optimized batch can be seen in Figure 3.The vesicle size was depicted to be 215.5 ± 2.25 nm which was very close to the predicted value with a % bias of just 0.93.The zeta potential of the optimized batch was found to be À 16.5 mV.A standard range for a stable particulate formulation is À 30 to þ30 mV.The negative value of zeta potential confirms that there is electrostatic repulsion between particles which avert aggregation between particles.The observed zeta potential confirms the stability of formulated polyherbal phytosomes as seen Figure 4. [36] 3.4.3.FTIR FTIR spectra can be seen in Figure 5. FTIR spectrum of ASEE shows characteristic peaks at 3271.9 cm À 1 (O-H stretching), 2924.87 cm À 1 and 2853.74cm À 1 (C-H stretching) Figure 5a.FTIR spectrum of CTEE gives characteristic peaks at 3273.11 cm À 1 (O-H stretching), 2975.71cm À 1 , 2853.74 cm À 1 (C-H stretching) and 1604.92cm À 1 (C ¼ C stretching) Figure 5b.The FTIR spectrum of soya lecithin exhibited the characteristic signals at 3370.55, 2923.18, and 2853.45cm À 1 , related to the C-H stretching present in the long fatty acid chain.Additional signals were also observed at 1735.43 cm À 1 (C ¼ O stretching in the fatty acid ester), 1243.54 cm À 1 (P ¼ O stretching), 1089.06 cm À 1 (P-O-C stretching), 1059.28 cm À 1 (C-N stretching) and 967.60 cm À 1 ([-Nþ (CH3]) in Figure 5c.
The spectra of cholesterol revealed the characteristic peak at 1362.52 (P ¼ O), 1507.96,11456.4cm À 1 (C-H vibration) and 1748.48 cm À 1 (C ¼ O stretching) in Figure 5d.The spectra of the physical mixture showed peaks that were characteristic of the individual components, that is, ASEE, CTEE, soya Lecithin and cholesterol at 3288.09, 2922.32,2852.32,1735.53,1612.64,1243.74 and 1049.11cm À 1 in Figure 5e.When we compare the physical mixture and the optimized PHP it is found that there is an increase in the intensity of H-bonding that would have led to complexation.There is a change in stretching frequency of O-H from 3288.09 to 3350.57cm À 1 in the optimized PHP which suggests the weak intermolecular interactions generated during formation of phytosomes.This is in agreement with earlier study. [37]A decrease in intensity of C-H stretching below 3000 cm À 1 and complete alteration around the absorption peaks of Soya lecithin at 1243.54 cm À 1 and 1089.06 cm À 1 indicates that there in an engagement of functional groups of Soya Lecithin in the formulated phytosomes in Figure 5f.

DSC
Differential scanning calorimetry is a quick, accurate way to look at the compatibility of drug excipients and how different components interact with one another.The various signs of possible interactions are emergence of a new peak, disappearance of an endothermic peak, change in peak form, change in onset temperature, change in melting point, difference in relative peak area or enthalpy. [38]Thermogram of ASEE, CTEE, Soya lecithin, Cholesterol, physical mixture and the prepared polyherbal phytosome complex are depicted in Figure 6.ASEE showed two endothermic peaks at 163.43 � C and 178.44 � C shown in Figure 6a.Similarly, CTEE also exhibited two endothermic peaks one at   176.40 � C and the other at 194.66 � C shown in Figure 6b.Soya lecithin showed its endothermic peaks at 130.10 � C and 276.40 � C shown in Figure 6c.The latter peak may be due the shuffling of hydrophobic tails of soya lecithin which leads to conversion of gel like structure to crystal liquid state.These findings agree with the earlier reported works. [39]hereas cholesterol showed one endothermic peak at 153.59 � C as seen in Figure 6d.The optimized PHP complex showed a single fused peak at around 148.84 � C which was found different from other peaks of individual components as seen in Figure 6e.Thus, it may be inferred that there may be an interaction between the extracts and Soya lecithin through van der Waals forces or hydrogen bonding, or the combination of both.This indicates the successful formation of the complex.

XRD
ASEE showed 2� value at 26 � , 51.6 � , 65.4 � and 74.8 � as seen in Figure 7a.whereas CTEE showed at 16.2 � , 25.6 � and 51.6 � as seen in Figure 7b.Both the plant extracts exhibited less intense and broader peaks.Soya lecithin showed 2� at 19.8 � as a big broad peak suggesting its amorphous character as shown in Figure 7c.This character of soya lecithin agrees with the previous findings also. [40,41]holesterol showed more intense peaks at 16.2 � and 17.2 � along with slightly less intense peaks at 14.2 � and 18.6 � which confirms its crystalline nature as shown in Figure 7d.However, if we look at the peaks of the optimized PHP the peaks of crude extracts, soya lecithin and cholesterol were not visible which suggests phase transformation. [42]The peaks of the PHP were significantly broader and less intense at 20.6 � (Figure 7e) which confirms the formation of complex in which the extracts are present in an amorphous state in the soya lecithin matrix. [34]4.6.SEM SEM images provide crucial information about the surface shape and solid-state characteristics.The SEM image of the optimized PHP revealed that they had a spheroid or irregular form and a rough surface texture as shown in Figure 8.

In vitro drug release
The release profile of the optimized PHP is compared with PHEE solution of polyherbal extracts in the ratio of 1:1.The comparative release profile can be seen in Figure 9.It was found that in case of PHEE 52.23% whereas in case of optimized PHP formulation a cumulative percentage release of 84.22% was seen in 24 hours.In case of PHP a slight burst release was also noticed within 2 hours of the study which may be due to deposition of drug on the phytosomes particles.A comparative fast release was seen up to 12 hours whereas a slight sustained and controlled effect in the release can be seen over the remaining time i.e., between 12 to 24 hours.This could be explained from the kinetic models.

Drug release kinetics
For estimating the release kinetics, the data is fitted in several kinetic models i.e., first-order, zero-order, Korsmeyer-Peppas and Higuchi models. [43]On the basis of highest value of correlation coefficient (R 2 ), the order of release was selected (S8).It can be seen from the table that the highest value of correlation coefficient was seen in Korsmeyer-Peppas models as 0.9817.On the basis of this the Korsmeyer-Peppas was found to be the best fit kinetic model explaining the release mechanism of the optimized PHP.The value of n was found between 0.45 and 0.89 (n ¼ 0.56).It suggests the release of drug from prepared phytosomes is non-Fickian diffusion.As per the Korsmeyer's equation the release mechanism could be diffusion and erosion. [44]

Development and optimization of phytosomal topical gel
The gels prepared from Carbopol 934 (PG1-PG3) were evaluated for physical appearance, pH, homogeneity, viscosity, and drug content.It was observed that the PG2 (1% Carbopol 934 gel) was found to be transparent.It had good homogeneity and excellent consistency.It was kept for 24 hours and no phase separation was observed.The details of optimization can be seen in Table 2.The optimized gel was then used for further characterization.

Ex vivo permeation studies
For an effective topical delivery of drugs, it is essential that the drug cross the stratum corneum and reaches the underneath layers.The rate and extent of percutaneous absorption influence the reach of the drug.The permeation of the drug from the Phytosomal gel was compared with a conventional gel by dispersing the PHEE in the Carbopol 934. [45]he permeation parameters are presented in S9.It was observed that 17.06 ± 1.54 mg/cm 2 of quercetin was permeated in 24 hours through the conventional gel in contrast 36.62 ± 3.33 mg/cm 2 of quercetin was permeated through the PG Figure 10 and S10.There was a significant (p < 0.05) improvement in the permeation of drug through the PG as compared to conventional gel.In case of conventional gel, the reason may be the poor permeation of phenolic and flavonoid compounds through the skin.It has also been reported in various findings that in case of conventional semisolid dosage forms the permeation of these types of compounds is poor. [46]he steady state flux (J ss ) for both the formulations was also determined.It was observed that the flux for the PG and the conventional gel was 4.05 ± 0.25 mg/cm 2 /h and 1.77 ± 0.42 mg/cm 2 /h respectively (S11).This increase in flux and permeation coefficient exhibits improved permeation from the PG as compared to conventional gel.The incorporation of PHEE into phytosomes enhances the permeation of herbal extracts through the skin.This study also agrees with previous findings. [29]The phospholipid present in the phytosomes helps in carrying the phenolic and flavonoid compounds across the lipophilic stratum corneum. [47]There is almost two-fold enhancement in permeation of quercetin from the PG which can be revealed by the    Enhancement ratio (E r ) of 2.14 ± 0.15.The increase in permeation might be caused by the formation of strong hydrogen bond which could have probably increased the absorption of phytoconstituents through the skin. [48]

Rheological behaviour
The role of viscosity is very important with regard to stability of the formulation and its application.It was found that the viscosity decreased upon an increase in the shear rate and vice versa which can be seen in the upwards and downward curves in Figure 11a, b.These curves demonstrate thixotropic behavior of the gels.The aforementioned rheological characteristics satisfied the formulation's absolute requirements for topical application. [49]A similar rheological profile of both PG2 and placebo gel was observed.This shows that the incorporated PHP does not affect the rheology of the PG2.

Texture analysis of the optimized gel PG2
The texture profile graph for the optimized gel PG2, which shows the relationship between force and time with the cycle of compression and decompression can be found in S12.The cycle has regular and smooth curve, which are a sign of the gel's smoothness.The hardness of the gel was found to be 78.4 ± 2.3 gm.The hardness is the parameter which characterizes the strength of gel structure when put under compression.It also conveys the ease of application of the gel on the skin.The adhesion on the other hand represents the bioadhesion which is required for retention of drug in the skin and a prolonged effect.The Adhesiveness was found to be 9.32 ± 1.23 mJ. [50]

Stability studies of optimized PHP and PG2
To ensure the stability of any formulation it is essential to perform its stability studies.Therefore, stability studies were performed.The details of the stability studies at three different storage conditions are given in Tables 3 and 4.There were no significant changes in the vesicle size and entrapment efficiency at refrigerated conditions (5 ± 3 � C), whereas at 25 ± 2 � C/60 ± 5% RH and at accelerated conditions (40 ± 2 � C/75 ± 5% RH) a slight deviation was observed in vesicle size.The slight increase in particle size may be due to aggregation of vesicles.However, there was not any significant change in the entrapment efficiency at 25 ± 2 � C/60 ± 5% RH but at higher temperature (40 ± 2 � C/75 ± 5% RH) due to partial degradation of layer of phospholipid the drug may have leaked which causes a decrease in entrapment efficiency. [51]he phytosomal gel was found to be stable during the study with no changes in physical appearance.There were no significant changes in drug content and viscosity during the storage

Conclusion
PHEE loaded phytosomes were successfully prepared using thin layer hydration method and optimized by CCD.These phytosomes yielded a high Entrapment efficiency.The vesicle size was in nanometer range with rough spherical shape.FTIR indicated the formation of complex between PHEE and Soya lecithin.DSC and XRD also support the formation of complex.The in vitro release from the phytosomes showed a biphasic behavior initial burst release followed by slow release.As per the R 2 value the best fitted model for release kinetics was Korsmeyer-Peppas model.These prepared phytosomes were successfully incorporated in the gel.The prepared gel was homogenous and with good consistency.The ex vivo permeation study revealed enhanced permeation of the drug across the skin with a significant enhancement ratio as compared to a conventional gel.The gel exhibited excellent rheological behavior required for topical application.As Annona squamosa and Cinnamomum tamala leaves have various reported activities like antimicrobial, antifungal, anti-inflammatory and wound healing which are related to their topical use.Some conventional topical dosage forms of Annona squamosa and Cinnamomum tamala have been reported in earlier research.But these formulations have certain limitation related to permeation.As already proved by this study the permeation of herbal extracts was found to be increased when complexed with phospholipids.Also, a polyherbal formulation by combining these two extracts has not been reported yet.A polyherbal formulation works on a holistic approach and exhibit a synergistic effect.Thus, it can be concluded that a Phytosomal

Figure 1 .
Figure 1.(a and b) 3-D response surface and contour plots showing relative effects of different independent variables on vesicle size.

Figure 2 .
Figure 2. (a and b) 3-D response surface and contour plots showing relative effects of different independent variables on entrapment efficiency.

Figure 3 .
Figure 3. Vesicle size of the optimized PHP.

Figure 4 .
Figure 4. Zeta potential of the optimized PHP.

Figure 8 .
Figure 8. SEM images of prepared PHP.

Figure 9 .
Figure 9. Comparative in vitro release of PHEE and optimized PHP.

Table 1 .
Response surface parameters of PHP formulation.
Note: All readings are in triplicate mean ± S.D.

Table 2 .
Optimization of the phytosomal topical gel.